https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116068
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |hubicka at gcc dot gnu.org
Ever confirmed|0 |1
Last reconfirmed| |2024-07-24
Status|UNCONFIRMED |NEW
Target Milestone|--- |15.0
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
The ICE is new (sanity checking), the issue is latent (and we might leak bitmap
memory this way).
IPA opts eventually using bitmaps need to initialize the global bitmap obstack.
I'm not sure what the reason was to not initialize that around IPA opts
done in process_new_functions?
Honza? Where's the best place to init/deinit for this backtrace?
